ChatGPT Medical Advice Lawsuit—What The Research Says About AI Diagnosis - Forbes

What the report says

Forbes reports that former Florida pastor Scott Winters has sued OpenAI and CEO Sam Altman in San Francisco County Superior Court, alleging that medical guidance from ChatGPT-4o in 2025 delayed treatment for a pulmonary embolism. According to the article, the July 2026 lawsuit says Winters repeatedly asked the chatbot about dizziness and unstable blood pressure, was reassured that the symptoms were not urgent and was advised to remain largely immobile. Weeks later, he suffered a serious clot in his lungs, and his legal team is seeking damages and an injunction pausing OpenAI’s health-focused ChatGPT feature pending outside safety review.

The article says OpenAI maintains that ChatGPT is not intended to replace a healthcare professional and that its terms warn users not to treat it as their only source of medical advice. Forbes also notes a separate lawsuit filed by a Texas couple after their son died of an overdose, framing the cases as part of a wider legal test over how much responsibility AI companies may bear when users seek help in medical or psychological crises.

Forbes contributor Jesse Pines, a physician, reviews peer-reviewed research showing a mixed picture. In structured diagnostic tests, some AI systems have matched or outperformed doctors, including studies in JAMA publications, Nature and npj Digital Medicine. But other research cited in the article found weaker performance under uncertainty, a tendency to accept fabricated clinical details, and potentially harmful omissions in medical recommendations.

The key distinction, according to the Forbes analysis, is between controlled research settings and unsupervised, open-ended conversations with incomplete patient information. The Winters allegations underscore concerns that chatbots may give overly reassuring or incomplete advice when a clinician’s examination, monitoring and judgment are needed.